Applying Machine Learning Algorithms for the Classification of Sleep Disorders

Abstract

Sleep-related disorders have a significant impact on both physical and mental well-being, highlighting the need for a reliable and accessible diagnostic method. While Polysomnography (PSG) remains the clinical benchmark for diagnosing sleep issues, it is often considered impractical due to high costs, discomfort, and limited availability. This project focuses on leveraging machine learning techniques to classify various sleep disorders using health and lifestyle data from the Kaggle Sleep Health and Lifestyle Dataset.Traditional approaches typically employ models such as K-Nearest Neighbors (KNN), Support Vector Machines (SVM), Decision Trees, Random Forests, and Artificial Neural Networks (ANN). Although effective, these methods are often computationally demanding and sensitive to hyperparameter tuning, which may affect their performance in real-world applications. To overcome these limitations, the proposed system utilizes ensemble learning techniques, specifically Stacking and Voting Classifiers, to enhance classification accuracy, stability, and model interpretability.By integrating the predictive strengths of multiple base models, the system aims to offer a more efficient, cost-effective, and user-friendly alternative to conventional diagnostic tools. Ultimately, this approach aspires to support early and accurate detection of sleep disorders like insomnia and sleep apnea, thereby improving patient outcomes and overall quality of life.